Python: [Bug]: Empty usage and finish_reason when a response is truncated at max_output_tokens #7051

import asyncio

from agent_framework import Agent, ChatOptions, Content, Message
from agent_framework_foundry import FoundryChatClient


async def main() -> None:
    client = FoundryChatClient(...)  # Azure AI Foundry project endpoint
    agent = Agent(client=client, name="TEST", instructions="You are helpful.")

    message = [
        Message(
            role="user",
            contents=[
                Content.from_text(
                    "Write an extremely long, detailed essay about the history of pizza."
                )
            ],
        )
    ]

    # Cap the output low so the model is truncated mid-answer.
    stream = agent.run(message, stream=True, options=ChatOptions(max_tokens=1000))
    async for _update in stream:
        pass

    final = await stream.get_final_response()

    print("text tail :", repr(final.text[-60:]))   # truncated mid-sentence
    print("usage     :", final.usage_details)       # -> None  (BUG: usage lost)
    print("finish    :", final.finish_reason)        # -> None  (BUG: no "length")


asyncio.run(main())